[REVIEW] Open Setup Reviews

[Jingle]

Travelling Desperados

6 VT
1 Role-Replicating 'Desperado'

3 Goon

If the Role Replicating 'Desperado' targets scum, the scum player dies. If they target town, they die and their target becomes a Role Replicating Desperado. Targets are submitted via PM.

[Jingle]

Day/Night would function normally. Each desperado would have a choice of trying to get their targets lynched normally or outing themselves to shoot. A desperado would functionally never be able to be lynched because they could always just shoot before the hammer, but they could be nightkilled.

If someone shoots scum they can't use the power again that day. A new desperado can use the power the same day though.

[Jingle]

Having now read the setup for popcorn mafia, I'd argue it's sufficiently different on the grounds that the mafia have no influence on who controls the daykills, town is still capable of lynching, and it's important for the Desperado to keep themselves hidden.

I kind of agree that the setup isn't very necessary as there is a similar one already existing, though.

[Jingle]

As written, the crossshoot endgame would be a townwin. Town doesn't need to have a player left alive to win.

And I'm fine with Mime winning Vig/Mafia/Mime despite the fact that the game could technically end in a different way because it's very unlikely to happen AND if it does the mime probably deserves the win for dodging multiple nightkills for however long it took to get to a 3p LYLO. If you let day start with 1v1v1 Mime cannot win, and it sucks to have a player who can't win in a game.

Also, it's impossible to have a 1v1v1 LYLO without a mime lynch. If a mime is shot, they both leave the game.

Also, A 0v2v1 should autoend in a mafia win because while mimes have majority they can't lynch both of themselves before the mafia nks. This should probably never happen, but I just realized it's technically possible.

[Jingle]

Huh. That should be fixed, because a player who cannot win remaining in the game shouldn't ever be a thing.

[Jingle]

Also, A 0v2v1 should autoend in a mafia win because while mimes have majority they can't lynch both of themselves before the mafia nks. This should probably never happen, but I just realized it's technically possible.

Nope, the Mafia can be continuously blocked by the Mime roleblock.

Yup. And lynch Mime #1, mafia have parity and win.

[Jingle]

Agreed. And 0v2v1 should be a mafia win, because the second mime lynch is impossible. 0v1v1 can only end in no lynch no kill cycles if the game is continued.

[Jingle]

Change to version where Mime must match death of other Mime. If Mime A is nightkilled, Mime B must be nightkilled.

If this isn't the fix, this actually might be a good basis for a discrete setup.

[Jingle]

Nope. Mafia lynch ends the game in town's favor. Mimes lose if they lynch the last mafia. God this is a wonky setup.

[Jingle]

Panic Room v2

11 players:

1 Mafia "Bomb" (Compulsive Multi-Vengeful)

2 Mafia Goons

1 Town "Recruiter" (Compulsive Day Neighborizer)

1 Town Recruit

6 Vanilla Townies

• Upon the death of the Mafia Bomb, everyone not in the Neighborhood is killed. After this point, the game becomes nightless.
  
• The recruit begins the game in the neighborhood along with the Recruiter.
  
• The most recent living recruit is always the one with the ability to recruit.

Anyone have issues with this being run in the open queue?

[Jingle]

Assuming activation is compulsive (It really should be):

The mafia can only kill if they've been activated thing is a win more mechanic. Odds that a mafia is activated D1 are ~96%. If no mafia are activated, you functionally have three conftown and the game is over.

Activators are irrelevant for EV, so it's probably functionally the same as 4v3.

You probably end up with a group decision as to who to lynch, so the activations don't matter as much except to mean mafia can't out themselves hammering. Mafia win condition has to be the homies dying or it's impossible to win with any mafia lynch. Which means mafia needs only two mislynches, possibly less. My guess without working out probabilities is that you need to add townies and get rid of the conditional mafia kill mechanic.

Assuming activation is non compulsive: You pick the towniest player D1 and that player's faction wins.

Edit: You can also call your activators treestumps. It means they can't be killed or vote.

[Jingle]

Extra kills are hard to balance, and not activating mafia means mafia have less influence the lynch. If you want to give value to sneaking onto the activation train I'd reward it with a 1-shot factional suicide bomber or a mechanic like forcing one of the activators into a specific choice. Or alternatively, you could make activated players lynchproof.

[Jingle]

There's a breaking strategy for town in Enter's setup.

First off, the Activators claim. As they're bulletproof, they have no reason not to. This removes them from the lynch pool.
Next, players, including Activators, unofficially vote (using FoSes or whatever) on a non-Activator player they believe is town. Then all the Activators choose that player as the player to activate.
Next, players, including Activators, unofficially vote (using FoSes or whatever) on a non-Activator player they believe is scum. The Activated player is forced to vote for that player, on pain of outing themself as scum. Likewise, if scum kill, this will out them as scum.
If an Activated player doesn't violate the breaking strategy, they should be chosen as the Activated player the following day, unless they're the desired lynch for the day.

Despite all that, EV is only 31.97%, because if scum is chosen on the first vote, their optimal strategy is to lynch the wrong player and nightkill another, even though they'll be outed in the process. Jingle wasn't kidding when he said extra kills were hard to balance.

For what it's worth, if you take the breaking strategy into account, the number of Activators is irrelevant, and the following ratios of homies:scum appear to be balanced (assuming I haven't made a mistake in the EV calculations or missed a better breaking strategy): 2:1; 3:2; 5:3; 7:4; 9:5. I'm not sure that the game would be much fun to play when broken, though.

I kinda said that, but props for doing the math man.

I kinda assumed activators were public, and to make the game more fun it's probably better if peeps can /in as activators/nonactivators. I think there's a lot of call for games in which there is a certain percentage of players who can ensure they're town.

Also 12:3 is probably too many homies. I'm gonna guess the EV is close to 7:3, (presumed 3.5-5 mislynches) but I'd appreciate if cfj would math for me cause I'm way too drunk to do probability tonight.

(also, it took 4 attempts ot spell probability.)

[Jingle]

For a viable EV calculation you probably assume scum kills whenever they have the option. It's not a perfect model (but then again, given subjective values there rarely is a perfect model) but it simplifies it enough that you can get a value. Honestly though, in this case it might be worth far more to just pick what seems to be a reasonable number and run it, then reevaluate based on results. Thinking over it again I would suggest 8 homies v 3 goons as a decent starting place. It's on odds and 8 v 3 is scumsided without PRs while the mechanic is probably very townsided from my expectations. Somewhere on par with an investigative doctor, afaict.

Still think you shouldn't make the NK dependent on activating scum, but :shrugs:

If you get enough people interested (5 pre-ins?) I'd either run it or comod it for you in either the open or theme queues, assuming you need an experienced mod to get it into a queue.

[Jingle]

Honestly feels less like a mafia game and more like a popularity contest. Any game where a big portion of it is picking who wins and loses within your team leaves a bad taste in my mouth.

[Jingle]

Interesting. I'm just going to assume that the opportunist doesn't know who the mafia are because it's broken if they do.

Balance wise, mafia is going to struggle to win this, but I don't think that's necessarily a problem. It's an interestingly viable strategy to no kill into 6p MYLO, because then even if town mislynches scum can choose not to kill which means they don't outright win.

You could probably make it a 7 person if you wanted a more even town/scumA/scumB win distribution. It'd functionally be 5v2 with an IC and nightless until the first scum dies, except, of course, that Opportunist becomes a Jester in 5p LYLO.

[Jingle]

As is it seems like there would be a pretty easy investigative loop, even discounting just being nightless.

Maybe put in a way to have scum stop coordinated actions? Something like if they guess all of a player's targets, that player doesn't act? probs not that, cause that seems like a nightmare for the scumteam, but :shrug:

[Jingle]

A>B
C>D
E>F
G>H

Doc claims N1.

Scum can't kill inside one of the pairs containing scum, so either they kill Conftown or they kill 2 town. If the scum who kills is in the targetee column, they're confscummed and a townie dies. If they're not they get to fake a result. So after night 1 it's 5p LYLO or 6 P MYLO with 2-3 cop results or faked cop results. Hm.

[Jingle]

Actually, Doctor Who being non targeting is interesting, and I don't think it hurts the setup.

Can scum win after a D1 scum lynch though?

[Jingle]

You have a 50/50 of a 1v1 for the player who doesn't kill, unless both scum are paired.
You have a 50/50 of the scum killing killing an extra townie unless both scum are paired, but that makes the killing scum confscum.

Hm... maybe copstart? That way the cops have to choose to target or not without coordination on N0.

[Jingle]

Maybe I'm misunderstanding, do the cops get town/scum when they successfully target?

Also, if the person targeting a player dies along with two others, it means that player acted. Which means they would be scum.

[Jingle]

Oh. Then it's probably not as unwinnable for scum if there's a scum lynch then.

[Jingle]

As yet unnamed jester night less variant:

3 Jester
2 mafia
4 town

Game ends when all members of a faction have been lynched, or # mafia >= # town.

Maf win if they have parity with the townies (discounting jesters).
Town win if all mafia are dead.
Jester win individually if they’re dead at endgame.

[Jingle]

Lynching 3 Jesters ends the game without hitting a mafia or town wincon, so if the Jesters claim one of them is guaranteed to lose. If only one Jester claims, the players just refuse to lynch that one player, and either town or mafia is guaranteed a win. If two Jesters claim, the other players can offer one of the two a win, but then the other is guaranteed to lose.

Definitely not worried about lynching Jesters.

Do you happen to have the 4:2 numbers available? I'm willing to bet the additional information from the Jesters helps town a little bit. Additionally, since it's a meme-harder variant of a meme setup, I'm willing to deal with a bit of imbalance in order to have a larger Jester percentage.

[Jingle]

Looked it up. According to an old Nightless EV formula from mith, 4:2 is 33% town EV. That's a really harsh winrate (especially if you take into account the chance of Only Jesters Win reducing that) but I think the additional length and the townsiding nature of nightless should help. Personally I think it's close enough to still be enjoyable, but if anyone can come up with a way to boost Town EV compared to scum w/out fucking with the Jester portion of the setup, I'd be interested.

[Jingle]

I disagree that the comparison is perfect. The inclusion of Jesters makes the game dramatically longer AND means scum has an incentive to not just lynch the weakest player. ime, the length of nightless games and the inability of scum to weed out strong voices is why nightless games townside over EV. That said, I agree with you that if there's a way to boost town v scum here that doesn't alter the core of the setup, that's good for the overall health of the setup.

Still doesn't really change the fact that what I'm looking for here is a boost to town that doesn't effect Jesters though. I'm not a big fan of IC's in nightless setups, tbh, but that is a potential solution. White Flag or variations thereof might work, but I feel that's a mite too punishing for Jesters. Adding players makes the game tedious, and removing mafia removes associations. I could see a vengekill for the first town lynch, but it seems inelegant. Investigations are largely the same.

[Jingle]

Not only did I just say that I wanted to avoid adding players because the game length is already a point of minor concern, that would bump the setup from a micro to a mini. I am 0% interested to wait through a mini queue for a meme setup designed to maximize the potential for a player who wants an opportunity to be a Jester to roll Jester.

[Jingle]

Alignment Cop seems iffy, given they're > IC. AC probably claims D1 (being IC adjacent isn't safe, it just means you get lynched from the other side) then activates in LYLO.

#s

11 -> 8 -> 5 -> Endgame

You have to average 1.3 scumlynches per day to win as town, and you have to lynch 2 scum in the first two days.

I assume 1v2v2 and 0v1v1 result in scumdraw, since in both situations town has lost and scum has no method of breaking a tiebreaker internally. 2v2v1 should result in larger scumteam win, because town can't win. (Team of two won't vote for an option that kills both of them, team of one won't vote for an option that kills them.)

[Jingle]

If the goal of the setup is to reduce the original Splash damage, I'd suggest reducing it out of multiball (a guess would be that 8v3 is more appropriate with just mafia choose list order). If the goal is to marathon-ize splash damage, it's already a fairly decent marathon setup if you can get the people for it. It's relatively short, there's no huge problem to meme-lynches and it's mechanically simple enough to pull off, assuming you think you can keep track of a 17p votecount in real time.

Ignore the math, I calculated awfully and don't have the time to fix it atm.

[Jingle]

/prein whatever setup gets named Splash Mountainous. I will likely retract this pre-in at some point in the future.

[Jingle]

Assuming mafia space themselves out to prevent double maf lynch is optimal (It might be? I'm not sure, but it seems like a reasonable strategy):

2/11 triple town D1.
Leads to 5v3
9/11 single scum.
Leads to 6v2

5v3
1/8 double scum into 3/5 LYLO
7/8 single scum into 1/5 LYLO
OR (Choice is scums)
3/8 double scum into 3/5 LYLO
1/4 Scumwin
3/8 Single Scum into 1/5 LYLO

6v2
1/4 Triple town into 3v2 LYLO, 1/5 town win.
3/4 1 Scum into 4v1 LYLO, 3/5 Town Win.

~45% Town EV. The difference in the scum choices is about 1% and it's the option which allows for an autowin for scum D2 which favors town, slightly.

Gonna go ahead and say it's theoretically balanced, unless someone thinks of a better strat for scum.

[Jingle]

Bb explains it well in the common mistakes thread, which is definitely worth a read. As far as balancing the setup, I could see three options. Total rework to be a semi open with variable roles, conversion to a small town, or transition to a closed setup, in which case you want to have this discussion in a pt.

[Jingle]

Now that I'm not mobile, I'll elaborate a little. Option one would be something along the lines of


Each of the following roles has a 33% chance of being included in this game. Any roles which are not included will instead by VTs.

Doctor
Tracker
Vengeful Townie
1-shot Governor
Hated Townie
Follower
Odd-night Commuter
Roleblocker
Borrower (Absorber, but only gets 1 shot for each absorption)
1-shot neighborizer

Smalltown is shorthand for a setup in which all of the roles are known, but the alignment of said roles is unknown. Thus you would have that list and three ability thieves, and randomly 3 among those players are scum.

Making a closed setup is pretty self explanatory, although you could either post a request for setup help in the looking for reviewers thread with your gimmick (scum action thiefs) or simply PM people you think might be willing/able to help you with balance.


You could also maybe change the setup fundamentally by making the town power roles repetitive (if there's 6 town trackers, scum can fakeclaim tracker) but doing so mean you need to have a very weak PR base and it seems pretty geared towards not having a bunch of the same PR as it is.

[Jingle]

Any particular reason you're starting on evens?

I think it's probably scumsided if boring if you use roleblocker. Probably balanced-ish with goon. Not sure why you're making that semi open.

[Jingle]

I am hesitant to add too many VTs, because even one or two weakens scum considerably

What?

Every pr drop comes with a town power drop that is > than the scum power drop.

Also, looking at this I'm fairly certain you're going to be more satisfied making a closed setup at this point anyway.

[Jingle]

Circle protect here is a terrible strategy, btw. It just outs the real doctor to scum if they get a save. A kill isn't incriminating, and a save still isn't a clear.

I get the idea of doc taking the game off of evens, but I think your doc save is significantly likely to only come late. (N1 doc save is 2/8*1/8+6/8*2/8 or ~22% assuming a scum lynch. It drops with a False Doc lynch and it drops dramatically with a Doc lynch.)

I don't think forcing a lynch is necessary at all. There's no reason that town wants to no lynch before MYLO or loss of all docs.

Similarly, I don't think 9v2 is unreasonable, given that town has no PRs who can avoid a lynch via claim.

[Jingle]

3v2v1 should endgame, actually. Dark Angel Claims, no lynch other than Dark Angel becomes possible, town can't win.

[Jingle]

My only issue with it is that mafia can get screwed by 3p playing poorly, but that's kind of the central conceit. Without working out EV I can't be sure it's balanced, but it's probably close enough that it's reasonable to run, and the players know what they're getting into.

[Jingle]

SK inside a scumteam is a thing that works, pops, although I'm not sure how much it really adds to the setup. I'd wager it ends up being significantly more townsided due to the lack of internal scum cohesion.

[Jingle]

That's also not to say it's not a valid game. There are 'mafia' games we play that are not really mafia, such as Dethy or low playercount AitP.

Scum should p much never shoot unless prs out, in which case one should qh and claim to be shooting that night. Thus, this is functionally 5v3 nightless. None of the setups seem like they'd have enough power to compensate for that, especially when you take into account that the game is functionally randomlynching. You could probably turn it into a decent enough large, but even still it'd be better run as a closed with little to no replay value.

[Jingle]

Monty Hall Mafia

3 Goon
6 VT

Plurality, standard nightkills.

At the halfway point of the dayphase, if there are fewer than three wagons, the lynch occurs immediately. Otherwise a random player who is being voted. If no player who is being voted can be confirmed to be town, Mafia immediately loses.

I haven't looked at balance much as this is a rough draft, but does anyone see any immediate concerns?

Edited to remove leashing as necessary.

[Jingle]

Mafia should attack revealed townie every night, making the reveal not very powerful.
After one mislynch, if town fails to build highiest wagon on mafia before midday, they lose.

I think you misunderstand the conceit of the setup. As long as there are 3 options, option 2 or 3 is revealed town halfway through the day. 3 options could be 7 votes for A, 1 vote for b, 1 vote for c.

The immediate lynch only triggers if there aren’t three options, which any individual group of three players can guarantee. Otherwise you continue the day with a conftown and theoretically it’s mathematically optimal to not lynch the person who had been the highest wagon (thus Monty hall).

I can dig the pops change though.

[Jingle]

God I hate conceited setups

Serious complaint?

Basically, the conceit of the setup is public cop halfway through each day randomly determined among all suspected parties. I'm really mostly concerned with fair balance. Town has more power than cop 9-er and it never goes away. Additionally, scum has the agency to shoot clears as soon as they pop up. 2v7 feels townsided and 3v6 seems scumsided, and that's where I'm most interested in feedback.

[Jingle]

Might try it as a 3v8, tbh.

[Jingle]

There being only two wagons is proscum as it means there is no cop, so not at all.

The concept isn't interesting enough for a large, imo.

Actually, given plurality and thus no real concern about being in evens, it might be a decent 2v6.

I'll probably run it sometime soon and see how it shakes out.

[Jingle]

Hypothetical EV doesn't matter at all in this case, as we can extrapolate practical information.

Vote for town is evidence (although weak) that the reversal of wincons doesn't change the observed winrate.

Black Flag Nightless is a roughly balanced setup via observation.

Therefore, reversed wincon Black Flag Nightless can be expected to be roughly balanced. It's probably more fun than the normal variant, too, since good play is rewarded with leaving the game so people who get demoralized are incentivized to try harder not lurk.

[Jingle]

Btw, any nightless setup changed to vote for town would have exactly opposite theoretical balance by necessity. Town wins in the cases it lost before, and scum wins in the cases it lost before.

[Jingle]

I would totally play as a guilty child in a Vote for Town setup. That sounds hilarious.

[Jingle]

First impression is scumsided.

Gut says assign a pool of players who are not allowed to act and a pool of players who investigate in the other pool. Second pool probably starts small (2-3) and increases to all but two on a scum lynch.

[Jingle]

Procrastination

1 Scum Mod-confirmed-role Beloved Princess
1 Scum 1-shot Watcher

1 Town Mod-confirmed-role Beloved Princess
2 Multi-acting Loyal Loud Friendly Neighbor Fruit Vendor
4 Vanilla Townies

What do you think?

Loud is redundant with FN. Night one, both FN's target both BP and one of them gets results, so scum is going to kill BPs early.

Lynch town D1, scum kills BP. Scum Kills not BP. Lynch Scum BP. Scum Kills 2x not BP.

Not taking into account the potential for shooting a FN, That's D3 MYLO. Watcher has a good chance of hitting one of if not both BPs. It's a 1/5 town win day one, functionally, and a MYLO with 0-2 conftown after that. Approximately a 47% town win if one of the FN's dies, otherwise a 60% town win. Doesn't seem untenable, but it's pretty punishing for town. It's an equilibrium on whether the optimal strats should be followed, and if they are it's functionally the same as a massclaim D1. Playable, but probably slightly scumsided.

[Jingle]

But it’s a 3 mafia, 4 town setup, where there’s 1 mafia goon, and 2 traitors, and the rest are all vanilla.
It’s nightless and if the goon gets lynched, the two traitors are endgamed. Lynching town loses day 1.

I don't think it meaningfully changes the math to drop a traitor and a VT (I mean, it does, but not enough to be insane) but it would reduce the PITA for the Goon if the traitors are lynched early. Interestingly, the traitors here are functionally identical to Unlynchers which might make the game more palatable on a reskin to some players. It's a scumsided by numbers (57% scumwin on D1) but not enough to invalidate the setup for what it is I think. Probably a good setup to run as a blitz or a marathon.

[Jingle]

Without running numbers it’s probably balanced. The only real issues I see is that it could get to functional autoloss but still have days left to get there and it’s stuck in perpetual evens. It might be more fun run as a 13p with White flag mechanics.

[Jingle]

Use a different green that one is physically painful.

[Jingle]

Alright.

Lone wolf is probably broken by repeated no lynching until puppet masters vig wolves.
Strings is roughly 2 goons v 2 mason doctors + 5vt which is still mafia but is townsided af. (The extra information mafia has over the masons is who is against them.)
SKD3 guardian angel doesn’t work as advertised and all of the setups seem scumsided at a glance. G angel just means LYLO is earlier because lynching the SK while purged outside of lylo is a guilty result.
DB is necessarily white flag, and there probably needs to be some kind of mechanic to prevent DB from switching prevents often or activity becomes AI. I’d suggest running it as a marathon, it seems fascinating that way. There’s also interesting wagonomics to be done with rvs, considering mafia bussing runs the risk of getting them day vigged. Prevented day kills and whether the maf pt has an access list do need to be clarified. If DB has access but isn’t mod announced they could just choose to be silent and force the mafia to hunt them in all of the players.

[Jingle]

Ironically my location changed just in time for me to be back in Spokane for several weeks. Hopefully I get to leave again in the next two days though.

[Jingle]

Nah, squatting in a house with limited utilities while setting it up to go on the market. Close though

[Jingle]

Is the accuse action public or private? If private, I'd suggest that DB doesn't need to be able to accuse, and it's honestly probably more interesting that way.

I dig the mafia into a game of magic train dynamic. It seems pretty awkward that 2 goons being lynched early means DB needs to pretend to be scum and still get lynched before endgame. It might be better to have it end 48 hours after the last mafia goon is lynched and prevent accuses from being changed once they've been cast. That way town doesn't have to theater a game of mafia where all the mafia are dead. OTOH, it would definitely be interesting to see two consecutive goon lynches and a low number of informed and watch DB accidentally endgame. In which case... What happens?

[Jingle]

Oh, also there should probably be a separate town and mafia dead pt, since correct play from town after being lynched is just to never post again which defeats the point of a dead thread.

[Jingle]

I'm thinking just to ask town players when confirming to given teh number of informs they'd like, take the median, and randomly distribute. Or I can ask Donnie his preference. Or I can just set it to be 3. decisions decisions.

Honestly, it probably doesn't meaningfully impact the game balance past whether it's known/unknown. There's upsides and downsides to both high and low numbers of informed, so any method of deciding should be fine.

I'd probably go with the first option to give players the illusion of being important though.

[Jingle]

I don't think you need to add any VTs, but I agree that the voting rule is a no go.

If 5p LYLO you can force 2 scummiest players to vote for 3rd scummiest player together to confirm they're not a scumteam. Only the two towniest players will be able to hammer. It's a risky confirm strat, but not one that should exist in the first place.

(Or a Mini Theme)

[Jingle]

I feel like the setup is townsided.

The first lynch either confirms an artifact as town or scum, and a JK means that the confirmed town artifact is unkillable or the confirmed scum artifact's lynch makes the JK a full cop. Similarly, the JK's partner is a mason. If the JK claims, mafia still only have a 50% to accurately target them.

tl;dr: there are a LOT of potential clears here. I do like the idea, but I don't think a protective role fits in the setup at all.

EDIT: removed a paragraph talking about unassigned artifacts, as that appears to have been a mistake on my part.

[Jingle]

There is no reason ever to lynch a guard or assassin for town, therefore you just out king/CL D1 and it’s literally a coin flip. CL can never win though, so putting forward any amount of effort is a solid king tell and town should functionally autowin.

[Jingle]

As stated in win condition: Town in order to win need to prevent assassin from winning.
As soon as the King/CL is revealed he will be inevitably killed(with nightkill or vengeful) and game will end in assassin victory.

I was assuming a CL lynch would end the game. If it doesn't, I'm not sure how this is anywhere near balanced anyway, tbh.

As it is, town's win con is lynch assassin in pool of 9 players without revealing the pool, then lynch all of the cultists if they somehow don't immediately lose. After only a few town deaths, cult is incredibly far ahead. If you have a large pool of targets Assassin and town are unlikely to win. If you have a small pool of targets, cult and town are unlikely to win. Anyway you look at it, town is just pretty fucked.

Optimal play for AitP is just shut up and lynch, and that benefits cult WAY more than town.

I'd wager it's functionally unplayable outside of marathons, but tbf I dislike AitP in the first place.

[Jingle]

Making Assassin only hit the named players and CL lynch a white flag might be enough of a boost, having given this a little more thought.

[Jingle]

IIRC, double day is one of the variants where the EV is solved. Can't remember where the thread with the math is, but a little bit of digging in mith's topics should turn it up.

With that said, 7v2 is probably the /most/ balanced of all the micro versions of it, but I doubt it's balanced or it would have a mention on the Double Day page. It's probably fine with a functional named townie though. This is based purely on my supposition that double day and a pseudo IC would be roughly equivalent to a cop and cop 9er being a balanced setup, rather than any attempt to find the EV calculations, so grain of salt etc.

[Jingle]

Aren’t Open Games usually 13 Peeps

Yes but that’s more a facet of laziness than a desired feature. 13p is easier to balance because the people who balance opens have experience with it.

If you come up with a setup that balances better with a different number than that’s actually desirable.

Generally speaking, you want an odd number of players for the majority of the game if possible and you should avoid setups that are prohibitively large (20+ people would be a nightmare to fill regularly) but non 13p setups (and setups that start at evens if it makes sense to) are actually more desirable than 13p imo because they add options to choose from.

[Jingle]

Balance wise, there’s no reason to need to force air to start as town. Otherwise it seems fine. Scum doesn’t have a nk, but they potentially have access to three killing roles. It’s a very skill based setup, but it seems fair. I’d personally leave the neighborhoods open during the day because I don’t see a reason not to but that’s a personal choice thing.

I’m assuming factional pt is a “neighborhood” but you could just leave it as a factional pt and give everyone a randomly assigned neighborhood to cut down on room for misunderstandings, since the factional pt isn’t really a neighborhood.

Darkness will be a struggle to keep people active, I’d bet.

Does air remove the player’s previous neighborhood ala fogport? Doesn’t really change much, but it would make for some interesting analysis.

Worth noting that you have 3kpn, which is high swing. Not sure how much you care about that.

[Jingle]

How are list ties resolved?

Could you clarify RC results probability? are they naive T/(T+M) amount of the time and sane the rest of the time? (That's how I parse it, but I think it's ambiguous enough that it should be clarified anyway.) I'm assuming it's not the traditional random cop where it just fails a certain percentage of the time, but sane/insane is probably less breakable than naive/sane.

[Jingle]

I feel like there might be strategy in massclaiming readlist submissions. If everyone picks a different mafia submission and puts the rest in unknown, then inaccurate flips can be used as pseudo clears.

[Jingle]

Eh. Bp cult leader should be enough. Vig doesn’t really make sense thematically, imo, but it should be balanced enough.

[Jingle]

Well, only having a fortnight to lynch everyone I see as a handicap. Isn't that a little

too

short of a deadline? So multi-voting and not having to replace votes helps gets lynches quicker.
Would it work better as a 5/12 large game?

FWIW votes not resetting at EOD creates some interesting quick hammer potential in the days leading up to XLO.

[Jingle]

Even with no-lynching at night effectively banned, I can't see any reason why the town wouldn't just direct the night vote onto a particular player, outnumbering the scum + their redirects in the process. Just like most scumsided voting mechanics, town can get back to regular voting mechanics via pseudovoting and then all massively piling on the same player.

Generally tactics like this being optimal make the game unfun.

[Jingle]

Technically speaking experienced mods can try to mod any open setup in then queue afaik. With that said, Tbone would probably interfere in the case of an excessively dumb setup and it’s always a good idea to get balance advice.

Also the change looks sound to me.

[Jingle]

Your EV calc is off, btw. When you factor in the repeated role elimination bit you get that mafia should win 5/9 games. (1/3 mafia is the last pick in the first pool plus 2/3*1/3 mafia is the last pick in the second pool.) So slightly scumsided from an EV standpoint. 44% odds of a town win is still playable though.

[Jingle]

Allowing Doc guesses just makes it Nightless with no alts.

Also, scum probably just caches confirmed alts until they can win all at once. They should get more kills per night as the game goes on and people are confirmed not to be specific players.

It's probably better balanced at 2v13, given that 2v11 mountainous is still scumsided in practice and the gimmick is inherently scumsided.

I'm assuming that the players in the game are public knowledge as opposed to a true secret alt game where players are force replaced if they alt slip, btw. Also relevant: do players flip with their alt information?

[Jingle]

Are you my Partner?

2 Goons
7 VT

Secret Alt Game (Non anonymous signups)

Game begins with no nightphases.

D1 a list of the players in the game is published. Scum is informed of the main, but not the alt, of their partner. Each day, scum may guess their partner's alt. If they are correct, they are told so. Once both scum have guessed their partner's alt or one scum is dead, they are given a PT where they may discuss the game and decide on a nightkill. At any time, a scum player may post

Enable: Night

in the scum PT. This will turn the game into a mountainous game (with daytalk).

[Jingle]

It seems likely to me that RB/Cop/amour is the correct option. It also seems likely that the town should no lim D1.

I’ll get back to you on that.

[Jingle]

Public Cop has a standard name, actually. It's called Publishing Cop.

It seems to me that if you make it to 6 VT's dead town probably just wins.

I don't see any broken combinations of roles though. The setup is too complex for simple EV calcs to be meaningful.

Edit: My issue is with last scum standing, not town longevity. My brain was being weird about numbers.

[Jingle]

Yeah, my brain was being weird about the numbers. The thing that I'm actually concerned about is still a viable concern though. Once scum is down to one member, it's almost impossible for them to come back.

It doesn't make it unviable, necessarily, but the setup very much supports landslide victories in one direction or the other.

[Jingle]

Additionally, town is def going to use their PRs early and that game is going to be a mountainous for the vast majority of it. It would probably be a viable chat or F2F setup if you had a large dedicated group and a good way of tracking votes though.

[Jingle]

Don't knock F2F giant mountainous setups, Gypyx. That's how I first started playing mafia and they're a ton of fun.

On another glance at the setup it's broken by massclaim, btw.

Massclaiming basically forces scum to commit to PR pools and every scum elimination creates conftown.

[Jingle]

Specifically in formats when dayphases last like 10 minutes. It's low commitment and leads to funny resolutions.

[Jingle]

Babysitter probably sits on alien most of the game.

Does killing the alien trigger the babysitter’s kill?

[Jingle]

IIRC the consensus is that games of that nature are better run in the Mishmash as Molla suggested. The departure being a lack of nightkill and day elimination.

Secret Hitler also falls under this category and is run semi regularly in the Mishmash afaik.

Seems balanced enough to run from what I remember of the Merlin version of Avalon, though, which is the closest comparison to your game I can think of. In fact, I think mechanically your game is a direct comparison to Avalon with Merlin and Oberon, while allowing for a voluntary passing of the team leader role that doesn't count towards the failed team limit, although I'd have to dig out the game to check the details on how many players go on each mission and how many failed teams lose the game.

[Jingle]

Babysitter probably sits on alien most of the game.

Does killing the alien trigger the babysitter’s kill?

haven't really outlined that yet but i'd say no

making the alien macho is an idea but then this would probably go into scum-sided territory and make this even more swingy

It already seems scumsided, tbh.

Town's only investigative role probably only nets a guilty on one scum on night one and gives unreliable clears, unless a protection is achieved. Meanwhile, scum has a role that helps them find the protectives and a role that means an outed protective is functionally useless, while town's real power is massively nerfed by being lovers. I'd suggest getting rid of the Encryptor and the MD. Alternatively, you could keep the encryptor but make the psychologist unlimited shot, but that turns town power into a win more proposition where the game gets exponentially harder for scum with each scum elimination, which is probably undesirable considering the Alien already does that.

[Jingle]

hmmm, alright, haven't heard of it named like that's cool though

i'll consider running it soon then, what's avalon?

Fantasy reskin of Resistance. It was the one that introduced variant roles past the all-knowing resistance member, and for a while was the only way to pick up a copy of the game. The roles are included in the article BB linked to.

FWIW, ETL also did a mafia/Resistance combination game at some point that was pretty cool, and if anyone wanted to redesign a setup like that or the aforementioned Chromavalon it would likely be well received.

[Jingle]

Now that I'm thinking about it, might be beneficial for Alien to claim day 1 to force a counter claim or even just to force the Roleblocker to roleblock them night 1 and 2.

You can do a weak follow the cop.

If alien claims D1 scum can MD an unclaimed town, kill an unclaimed town for a 1/9 chance of killing both and a 1/9 of catching a town PR. Town gets a 0% of stopping a kill, and it's exactly as likely for the psychologist to hit the babysitter as the scum who didn't make the kill, meaning a fake guilty is the same as an actual guilty and the Psych becomes a glorified IC. It also gives the scum the ability to narrow down the babysitter based on JK's actions.

Probably better if the Psych claims D2 without outing their target and the JK targets them at night so that you can play pseudo Follow the IC, as the psychologist is unlikely to die before scum kill the rest of the power because they won't be able to tell if they detected the JK or the babysitter.

Still think it's probably scum sided as is, tbh.

[Jingle]

1. is it fine from a balance perspective to have both a false negative AND a false positive here for psych in terms of alignment?
2. is it customary on mafiascum to tell PRs that their action failed when they target a macho player?

From a balance perspective? It's just two nerfs to the role, which is completely legitimate. The question of course being whether those two nerfs are necessary. It's a fine choice so long as the rest of the setup makes it make sense, especially in an open where it's known that those nerfs exist. In a closed setup, excessive nerfing of a role can be frustrating, but this particularly set (a fake inno and a fake guilty on a questionable reliability investigative) wouldn't be unreasonable.

Generally speaking, power roles aren't notified their action failed unless they get a visible result. For example, a tracker might receive a "No Result" PM because their role normally gives them a result PM. A doctor on the other hand would have no idea that their save failed because they wouldn't normally get a result PM. Of course, it's completely reasonable to have a setup where a doctor would be notified that their role failed, that would just have to be accounted for in the design.

[Jingle]

I think scum is more likely to find the masonry than in F&E, but I legitimately hadn't considered that as a starting point for the comparison. I do like the Macho Psych suggestion, and I think it solves the grossness of a follow the IC strat pretty well. I like that only one member of the masonry is ever safe from the NK, and I don't think it's likely that town ever nets a miselim or a psych guilty, but that the possibility for it to happen is a cool feature of the setup.

Depending on playerlist I'd probably play in the setup as described in 1116.

[Jingle]

Also, town tends to overperform in nightless setups.

I’d be interested to see the math if the first town is a vengecop and the first scum is a vengekill. Might make it scumsided enough. Not gonna do the math while mobile though.

[Jingle]

Is townplay onsite strong enough yet to just run 4:2 nightless? Lol.

Honestly? It might be. 6v2 is perfect EV, and towns do tend to do better than site meta in nightless games. I'd probably be more willing to play in a 4:2.

[Jingle]

Town only gets 1 miselim though, and then you are forced to guess right. I don't think the "town does well in nightless" applies as you are getting so little info in comparison to a longer nightless game.

That only applies to the case when scum get a perfect win (40% of games, tbf, but not all of the scum wins). If scum is ever eliminated, town nets another miselim.

EV on a vanilla 2:4 nightless is 67% scum win, which is undoubtedly high, but the lack of ability for scum to kill off particularly dangerous town voices is a large portion of the conventional wisdom on nightless setups being townsided.

I'm not saying it's a perfectly balanced game, but I would definitely be more interested in seeing it than a 2:6.

2:4 nightless, scum wins at two town left, if all executions are town, the last executed player may try to guess the scumteam for the win

random idea, how would it be?

Only changes the above percentages in the case of a scum sweep, mathematically a hail mary guess and an elimination are identical if there's only one scum remaining. The new EV is 56.6% swumwin. I would wager that being able to just elim the strongest scumhunter on the first game-loseable elim would further skew results towards townwins based on subjective speculation, but it's within the realm of testable for sure.

PEdit: didn't realize the guess wouldn't happen if a scum elim had happened. It's actually 70% scumwin as you suggested it.

[Jingle]

I think 2:5 Nightless with 2 Vengeful Mafia Goons would be fun

Mathematically identical to 2:4 nightless with less of a townsiding bias introduced by the inability to kill. Probably better at 2:6, but you'd have to run the setup a few times to be sure. Doesn't seem like a bad setup to try a few times either way though.

[Jingle]

Yes. Only one half of the masonry can be fully protected so it's a WIFOM game.

Definitely looks playable, although it might benefit from tweaks after being run once.

[Jingle]

I mean... you kinda have to run it in order to do tweaks after running it. I was telling you it looks fine as is.

[Jingle]

Subject: Mafia Are Masons

2 Mafia A Goons
2 Mafia B Goons
4 VT

During Night 1, the first player to be eliminated chooses either Mafia A or Mafia B to be recruited to the town as a Mason Pair. The day 1 yeet is compulsive and defaults to plurality if no majority is reached. Standard yeet rules apply on future days.

Notes: Both Mafia groups will have access to a nightkill on N1, but neither will know whether they are going to be town or mafia at the end of the night.

[Jingle]

Responded-ish in setup thread, but at least the make it harder for the scum to aim their kill bit is wrong. Both scumteams should be shooting for the other team, so discussion/scumhunting is a net positive on D1.

[Jingle]

There’s also the problem that a scum day vig is disproportionately more powerful than a town day vig and town power is very likely to be obvious when they buy.

Also cross targeting JKs is kinda icky.

I think your game is going to be a lot more vanilla than intended.

You could throw in a fairly inexpensive achievement cop and an expensive investigative role (watcher?) but as is I think you’re just heading towards a mostly mountainous game.

[Jingle]

achievement unlocked game

I uh... Don't remember that tbh.

I'd probably be willing to play this as a 13:4 with the right list but I'm honestly not convinced it's balanced. There's a certain value to knowing what you're getting yourself into that makes me more willing to try weird departures from the norm like this though.

[Jingle]

(your achievement unlocked game is what inspired this if it wasn't obvious btw)

I looked it up, this was a ducky and S_S game. I'm flattered that you'd assume it was me, but I'm not solely responsible for all of the weird mutations on site.

[Jingle]

Isis's scumgame so good she doesn't even have to ban the players who scumread her to win. Agreed that watcher below 4 would be too cheap. You could maybe play around with a mechanic to allow the sharing of Achievement Points depending on how likely it is all the achievements actually happen. The less points in play the more reasonable that option becomes.

[Jingle]

How do you deal with sequencing?

Like, if I set up a situation where a vote hammered someone which then hammered someone else which then hammered someone else, would the hammers all resolve and then the venges or would the venges resolve between each vote count check?

[Jingle]

^Agreed. EV is a bad metric for nightless vengeful but I'd probably want 35-45%.

original crossfire had only the first hammer resolve, no simultaneous stuff

Votes are not reset after an elim occurs (so combo elims can occur if someone else was on L-1 with the person who was eliminated not voting for that person).

[Jingle]

But that isn't the question?

VoteCount

A: AD
B: BD
C: D
D:

4 Alive, 3 to lim.

VOTE: A

A dies, now B is at majority. If venge first, A can kill D and save B. If lim first, B dies and only C and D are alive. It gets even wonkier if you add in a situation where C votes A and B.

PEdit: Thanks. What if C voted both A and B in my above example?

[Jingle]

My intention here is to determine if it's possible to force a simultaneous killing of all living slots and thus a townwin, btw.

[Jingle]

Are vengekills compulsive?

[Jingle]

My guess is that setup is scumsided, MC, but it's also probably fast enough that if you run it as a blitz/marathon people won't care that it's scumsided.

[Jingle]

I'm ok with a scum winrate of like 60%, you think it's in that ballpark or much worse?

The issue is that based on the impact of scum choosing # of passes we don't really know how the setup will pan out. There's a bit of play in that metric for how much it benefits town or scum and without seeing it be played all I can give you is gut instinct. I'd guess it's somewhere in the ballpark of 60%, but I honestly don't know a setup similar enough to draw a parallel to, so you kind of have to run it to find out.

With that said, experimental blitz games are usually pretty enjoyable, so balance is kind of a secondary concern, and I think the answer to "Is it likely to be fun" is yes.

[Jingle]

Balance relies a bit to much on how good people are at obfuscating alts, but it looks decent to me assuming at least some level of success.

[Jingle]

Secret alt games are not universally designed, but more often than not there's a drawback to getting correctly identified.

PEdit: Yeah, that. I've seen NKs that only work on identified players and factions outright winning if they can guess all the identities as well. And some just force replace anyone who main slips.

[Jingle]

Kind of meme-y mountainous setup I intend to run in the nearish future. Would appreciate feedback.

https://forum.mafiascum.net/viewtopic.php?f=115&t=88370

[Jingle]

(I'm also not even sure that 2 goons/4 VT/1 cop is even good for town. I honestly think 2 goons/5 VT/2 named townies is probably better EV.)

Technically it’s neither of those. It’s slightly more power than either case. In one you have an 8 person with a Nightstart (slightly more likely to get an investigate) and in the other it’s got the odds of a cop becoming apparent.

Regardless to get EV you’d run the calculations on both models and pick the higher one, but I’m not convinced it’s a setup where EV is the right approach to balance.

[Jingle]

More that the nuance on making a system for the latter involves a lot of branching paths which both complicates EV determination and makes it less reliable because of swing.

I’d look for double named townie runs and slightly course correct for additional townsidedness. Gut says playable but leans slightly town overall. Two clears on D2 of a micro is not insignificant.

[Jingle]

You could go with a “next post is a deputy claim style thing”, which could itself come from a deputy, but that seems tedious at best.

Honestly, regardless of EV I think it would just be more fun to play it as a 2 named town with an emergent cop setup and rely on day play.

[Jingle]

Technically not mafia, despite being hidden information (there's no informed minority). EV is exactly 50% if you change it to 3:3:3 and probably plays out similarly to BoP Mafia. (50/50 in each hood so all right and all wrong are equally likely, 50/50 in the final trio). There is a slight townsiding influence in that the IC in the second round knows at least one hood chose wrong, but it's not really useful from a POE stance because they don't know which one or whether said player is even in the final grouping.

For reference, BoP is observably balanced, but has a third townsiding faction instead of the ICs.

Notably: you could have a separate queue for the IC roles to allow players to self select whether they want to be in that role.

Also notably: the EV remains unchanged for every setup where the size is n:n:n.